It was raining the way it only rains in Hong Kong in late June — not falling so much as hanging in the air between buildings — when a friend in Kennedy Town sent me a screenshot of his PayMe screen. Between the usual tiles for splitting a dinner bill and topping up a card, there was a new one. RedCoin. He asked me what it was. I had spent the previous nine weeks reading whitepapers from projects that had each raised a hundred million dollars, and I could not answer him — because the entire public record of the most consequential stablecoin launch of the quarter fits into fewer words than a hackathon deck.
Here is everything actually established: HSBC has issued a Hong Kong dollar stablecoin called RedCoin; PayMe users can activate a stablecoin wallet inside the app; it is built for person-to-person and person-to-merchant payments; and it goes live for retail customers within days. That is the whole of it. No chain. No reserve composition. No licence number. No fee schedule. No float policy. No audit report. When a systemically important launch arrives with this little surface area, the omissions are themselves the analysis — what a bank chooses not to say tells you more than what it chooses to market.
I want to be careful about the register of this piece, because the temptation with bank stablecoins is either euphoria ("institutions are here") or contempt ("permissioned junk"). Both are lazy. What follows is what I can defend: a reading of the architecture that RedCoin implies, the reserve economics that actually pay for it, and the one structural question — stablecoin or tokenized deposit? — that almost nobody is asking and that changes the legal character of every dollar a user hands over.
The context that matters
PayMe is not a startup product. HSBC launched it in 2015 as a peer-to-peer transfer app tied to card top-ups, and over a decade it has become one of the few pieces of consumer software in Hong Kong with genuine mass reach — most public estimates put registered users in the low millions, a meaningful fraction of a city of 7.5 million. That installed base is the entire strategic premise. RedCoin is not trying to win crypto users; it is trying to convert an existing payment habit into a balance-holding habit.
The regulatory frame arrived first. Hong Kong's stablecoin regime, in force since mid-2025, requires any issuer of a fiat-referenced stablecoin targeting the local public to hold a licence from the HKMA, maintain full backing in high-quality liquid assets, segregate custody, submit to regular audit, and honour redemption at par. Ahead of that, the HKMA ran a sandbox whose participants were mostly consortiums — Standard Chartered with Animoca and HKT, JD's CoinLink — not the incumbent retail banks. HSBC's parallel track was quieter and, in my reading, more serious: years of participation in wholesale CBDC pilots, the Ensemble tokenized-deposit work, and the Orion bond experiments. A bank does not spend five years in tokenization pilots to ship a payments gimmick. It does so because it believes the settlement layer of its own balance sheet is about to be rewritten.
I have spent the last two years inside this machinery, comparing twelve CBDC prototypes and interviewing developers in Lisbon and Singapore about how they refactored contracts to survive the new rules. The lesson I keep returning to is that compliance is a design material, not a constraint — and that the banks which treat it as material build products that feel inevitable, while the ones that treat it as a tax build products that feel like forms. RedCoin, on the evidence available, belongs to the first category. Which is precisely why the missing details matter so much.
Distribution is the product, and that is not an insult
Let me say the unglamorous thing first: there is no technical novelty here that a competent protocol engineer would call novel. Fiat-backed issuance, custodial wallets, P2P and merchant payments — this is a solved design space. But solving a design space is not the bottleneck. The bottleneck for every crypto-native stablecoin has always been the same: acquiring real humans who have a reason to open the app on a Tuesday. Circle spent a decade and a public listing getting there. RedCoin starts with the humans already inside the wall.
That is a distribution moat measured not in TPS but in thumb-reach — one tile, one tap, one identity already verified. In my vocabulary, this is a flow innovation. The technology is ordinary; the placement is decisive. And placement is exactly the kind of advantage that crypto-native teams systematically underprice, because it does not appear in a whitepaper and cannot be forked.
The settlement layer is a black box, and the black box is the point
Now the part that should trouble anyone who cares about verifiability. The underlying ledger is undisclosed. Given the institution, the customer base, and the regulatory posture, I would put high confidence on RedCoin running on a permissioned distributed ledger or a gated bridge onto a public chain — essentially certainly not an open, permissionless network that anonymous contracts can compose against. The phrase "activate a stablecoin wallet inside PayMe" is doing a lot of work too: it describes custody. Keys, balances, and transfer permissions live on the bank's side of the glass.
I have no interest in moralizing about that. Custody is the correct design for a retail product regulated as a bank liability, and self-custody would be a customer-protection failure dressed as ideology. But the consequences are concrete: composability is approximately zero. RedCoin cannot be dropped into an automated market maker, cannot collateralize a loan on a lending protocol, cannot be recycled through a yield aggregator, and almost certainly cannot be bridged except to whitelisted addresses. The stablecoin multiplier — the mechanism by which one unit of reserve supports several units of on-chain activity as it recycles through lending markets — does not exist here. A reserve backs one unit, that unit moves once, and the loop closes.
RedCoin's developers are not working on-chain. They are working on a ledger with an auditor, an administrator, a freeze function, and a reconciliation job. I have audited enough of these to know what the operations runbook looks like, and I know that in such a system the interesting security surface is not the consensus algorithm. It is the role-based access control and the person who approves the whitelist.
Where the money actually comes from
Here is the analysis I think the market is missing entirely. RedCoin has no token economics in the conventional sense — it does not appreciate, it does not pay a dividend, it confers no governance. So the correct frame is not tokenomics but reserve economics: how does the issuer monetize a zero-margin payment instrument?
Three revenue lines exist. Merchant acceptance fees on the P2M side, which are real but competitive and will be compressed by Octopus, AlipayHK, and card rails. Deposit retention, which is defensive rather than additive — money staying inside HSBC rather than leaving. And float. Float is the quietest yield in banking, and it is almost always the real answer.
Run the numbers the way a treasury committee would. A user converts Hong Kong dollars into RedCoin; HSBC holds the corresponding dollars as a reserve; the reserve earns interest in the interbank market. If RedCoin reaches HK$5 billion in circulation — an optimistic but not absurd figure for a bank with this retail base — and the reserve earns roughly 3 percent, that is about HK$150 million a year of income on liabilities that pay the customer nothing. No credit risk beyond the sovereign, no duration mismatch worth naming, no marketing spend beyond a tile. That is the business. A transaction is just a promise frozen in time, and the issuer of the promise collects the interest while it thaws.
Except — and this is the part I have not seen anyone write — that annuity is hostage to the mechanics of the peg. In May 2025, when the HKMA intervened on the strong side of the convertibility band and the aggregate balance swelled, one-month HIBOR collapsed from around 4 percent to below 1 percent in a matter of days. The float income on a HK$5 billion book did not decline; it more or less evaporated, from roughly HK$150 million annualized to something in the neighbourhood of HK$40 million. RedCoin's profit centre is not a business model so much as a byproduct of where the currency sits inside its band — a monetary-policy residual that can be halved by a week of inflows. That is a strange foundation for a strategic product, and it explains why the bank will care far more about deposit retention than about float.
The negative-carry problem nobody wants to price
Follow the same arithmetic from the user's side and the design starts to look odd. Hong Kong dollar deposits pay interest. A non-yielding HKD token does not. At 3 percent, holding HK$10,000 in RedCoin for a month costs the user roughly HK$25 in foregone interest; at 0.8 percent it costs about HK$7. Small per person, but non-zero, and permanently one-directional. Why would a rational retail customer accept negative carry on a payment balance?
Because the carry is not the point. The point is the payee. A small merchant doing HK$200,000 a month in takings pays somewhere between 1.5 and 2.5 percent to card acquirers — call it HK$3,600. At a plausible 0.6 percent stablecoin acceptance cost, that becomes HK$1,200, a saving of HK$2,400 a month, or nearly HK$29,000 a year, against the customer's HK$300 of foregone interest on a HK$10,000 balance. The economics of RedCoin close on the merchant side of the transaction, not the consumer side. The payer subsidizes the payee. That asymmetry has a prediction attached to it: expect merchant-funded rebates and acceptance incentives, not bank-funded consumer rewards, because the surplus sits with the shopkeeper and the bank knows it.
The cold-start problem is bilateral and boring
Which brings us to the least glamorous risk in the whole file. A payment network needs two sides, and the merchant side is not built by shipping an app. It is built by salespeople signing acceptance agreements, one category at a time, in a city where Octopus owns transit and convenience retail, AlipayHK and WeChat Pay HK own cross-border shopping, and cards own everything else. No merchant count has been disclosed. That silence is louder than any technical gap.
Here is the test I will apply in ninety days: not wallet activations, which will look impressive because they are one tap, but the number of distinct merchants that have processed more than a thousand dollars of RedCoin volume. That single metric separates a product from a KPI ornament, and it is the number that will not appear in a press release.
The peg makes the de-dollarization story a fiction
I want to dispatch a narrative that has already begun attaching itself to this launch. Hong Kong dollar stablecoins are routinely described as instruments of currency diversification, a hedge against dollar hegemony, a step toward multipolar settlement. Under a linked exchange rate system that holds the currency inside a 7.75 to 7.85 band against the US dollar, the Hong Kong dollar is a shadow of the dollar, and a token redeemable one-for-one into it is a shadow of a shadow. RedCoin is, in monetary substance, a dollar-adjacent instrument with a different symbol on the front.
That is not a criticism; it is a clarification. The honest description of this product is local payments digitization — a modernisation of the retail rail — not monetary sovereignty. Anyone selling it as the latter is selling something the peg does not permit.
Reading the flow the way I read the twelve prototypes
When I compared CBDC prototypes across jurisdictions, the failures were almost never cryptographic. They were cognitive. A wallet that requires a user to understand the difference between a balance, a reserve, and a redemption right has already lost. RedCoin's flow — activate, top up, pay, receive — is short, and short is the entire achievement. Two taps between intent and settlement is the kind of design decision that gets ignored in technical reviews and decides adoption entirely.
But the same flow contains the omission. There is no visible interest accrual, no visible reserve attestation, no visible chain. The user is asked to extend trust and not to verify anything, and the interface is deliberately smooth about it. The chain asks you to verify. The bank asks you to trust. RedCoin asks the second and quietly implies the first.
The question nobody is asking: stablecoin or tokenized deposit?
Here is where I will stake my most contrarian technical claim. The word "stablecoin" may be doing marketing work for something legally different. A licensed bank issuing a digital claim on itself, redeemable at par from its own balance sheet, is structurally a tokenized deposit — and under most supervisory frameworks, banks can issue tokenized deposits under existing banking authority rather than under a separate stablecoin issuance licence.
If that is the structure, then the protections the stablecoin regime was built to guarantee — full segregation of reserve assets, statutory redemption rights, insolvency remoteness — may not apply in the same form, because the holder's claim is on the bank, not on a ring-fenced pool. In ordinary conditions this distinction is invisible. In a resolution scenario it is the only thing that matters. I have not seen HSBC's characterization of the instrument, and until I do, I treat every description of RedCoin's backing as marketing until proven otherwise. A ledger is not a ledger when the only auditor is the issuer.
None of this is an accusation. Banks are entitled to issue deposits; that is their business. But the crypto industry has spent a decade learning to read the difference between a claim on a pool and a claim on an institution, and that literacy should not be switched off because the issuer has a famous logo and a marble lobby.
The contrarian read: the competitor is the deposit account
The consensus framing is that HSBC has entered the stablecoin wars against Tether and Circle. I think that framing is wrong in a way that leads to bad forecasts. RedCoin is not competing with USDT, which serves users Hong Kong's banks will never onboard, nor with USDC, which serves developers who will never be able to compose with this product. It is competing with the Hong Kong dollar savings account and the Octopus card — with the passive balance on a customer's own statement, and with the stored-value card in their pocket.
The blind spot follows from that. Analysts will track RedCoin as a crypto asset and grade it on circulating supply; the bank will track it as a deposit franchise and grade it on balances retained and cost of acquisition avoided. Those two scorecards will diverge wildly, and the divergence will be read by the crypto side as failure. It will not be failure. It will be a different game being played on a different field with a different scoreboard.
And the centralization that crypto reviewers will list as a red flag — freeze functions, admin keys, whitelisted withdrawals — is not a flaw in this product. It is the product. The reason a shopkeeper in Sham Shui Po will accept a RedCoin payment and will not accept a USDT transfer is precisely that reversibility and recourse exist. What reads as a security weakness on a protocol forum reads as consumer protection in a wet market. Both readings are correct, and they are correct about different things.
What I am watching
Three numbers will tell me within two quarters whether this is infrastructure or ornament. The count of genuinely active merchants, measured in volume rather than logos. The redemption service level — how long a customer waits to convert back to a bank balance, and whether the terms say "at par" in a way a regulator would enforce. And the interoperability decision at the HKMA: if the authority mandates a common clearing layer for Hong Kong dollar tokens, HSBC's first-mover advantage in distribution gets socialized into public infrastructure, and the moat narrows overnight.
My suspicion, held loosely, is that the stablecoin is not the destination. It is the test load for a settlement layer the bank intends to use for tokenized deposits, tokenized bonds, and fund units — the place where the real margin lives. Retail payments are how you find out whether the plumbing survives contact with ordinary people.
So the question I will leave with the friend in Kennedy Town, and with anyone else who taps that new tile: when you convert your dollars into a promise, who is holding the interest while the promise waits, and what exactly do you own if the promiser stops promising? The interface will never ask you that. It is designed not to.
